With development of an Internet technology, Internet information increases explosively. If users can be divided into groups according to interests of the users, and information corresponding to an information category in which users are interested can be recommended for the users in a same group, a click-through rate of the information can be greatly improved; for example, if an information category in which users in a group are interested is casual clothes, a server recommends commodity information of the casual clothes for the users in the group. Therefore, how to acquire an information category in which a user is interested appears to be particularly important.
Currently, an information category in which a user is interested is mainly acquired by using a search engine, including: acquiring keywords entered by the user into the search engine within a period of time, and for each keyword entered by the user, calculating a matching degree between this keyword entered by the user and each keyword in a correspondence between an information category and a keyword, selecting a keyword whose matching degree with this keyword entered by the user is the maximum, and increasing search times of an information category corresponding to the selected keyword by one; and acquiring a preset quantity of information categories whose search times are the maximum within this period of time, and using the acquired information categories as information categories in which the user is interested.
During implementation of the present disclosure, the inventor discovers that the existing technology has at least the following problems:
For most users, a frequency of using a search engine is low, and a quantity of keywords entered into the search engine within a period of time is not many, and is even less than a preset quantity, in this way causing accuracy of an information category in which a user is interested, which is obtained according to a keyword entered by the user, to be low.